A conventional semiconductor light emitting element includes a double heterostructure that includes a light emitting (‘active’) layer that is sandwiched between an N-type clad layer and a P-type clad layer. When charge-carriers (electrons and holes) flow into the active layer, these charge-carriers may meet. When an electron meets a hole, it falls into a lower energy level, and releases energy in the form of a photon. The created photon may travel in any direction, and commercially available light emitting elements typically include reflective surfaces that redirect light so as to exit an intended escape surface of the light emitting element. However, the light may strike the escape surface at virtually any angle, and a substantial portion of the light may strike the surface at an angle that exceeds a critical angle of the interface between the materials on either side of the surface and be totally internally reflected (TIR).
The critical angle is determined by the indices of refraction n1 and n2 of the material at an interface between the materials, and is equal to:arcsin(n2/n1),  (Equation 1)for light traveling from the medium having an index of refraction n1 into a medium having an lower index of refraction of n2. Light that strikes the surface at greater than the critical angle will be totally internally reflected, and will not escape through the surface. The term “escape zone”, or “escape cone” is used to define the range of angles within which light will escape through the surface. The escape cone at any point on the surface is a cone with an apex at the surface whose cross-section subtends an angle of twice the critical angle about a normal to the surface. The escape zone is the composite of the escape cones of all points on the surface.
Although escape zones are defined by solid angles, this disclosure is presented using a two-dimension model, for ease of presentation and understanding. One of skill in the art will recognize that the conclusions drawn from the following analysis of two dimensional optical models are applicable to a more complex analysis using a three-dimensional model.
It has been determined that roughening the escape surface allows more light to escape through the surface, compared to a smooth surface. When light is totally internally reflected from the smooth escape surface, it will travel back toward the interior of the light emitting element, be reflected by the reflective surfaces, and redirected back toward the escape surface. In most cases, this process is repeated until the reflected light is fully absorbed inside the LED. Conversely, because a roughened surface will have portions of its surface at varying angles relative to the surface of the active area, some of the light that would have been outside the escape zone of a smooth escape surface will be within the escape zone of a sloped surface of the roughened surface and will exit the roughened escape surface; additionally, some of the light that may be reflected from the roughened escape surfaces may be redirected in the desired direction (e.g. orthogonal to the active layer), so that on the next bounce, the likelihood of exiting the escape surface is increased.
The escape surface of the light emitting element may be roughened using any of a variety of techniques, some of which create a randomly roughened surface, and some of which create a surface with a particular pattern of grooves, crevices, structures, and the like. In “Recent Progress of GaN Based High Power LED” (14th Optoelectronics and Communication Conference, 2009), Hao-chung Kuo discloses a combination of roughening techniques wherein the escape surface is first patterned, then subjected to a random roughening process, creating an escape surface having a roughened pattern.
Although roughening the escape surface improves the light extraction efficiency, some of this efficiency is lost when light that exits a feature on the roughened escape surface strikes an adjacent feature and is ‘re-injected’ into the light emitting element. Additionally, in a random roughening process, control of the shape and density of the created features is somewhat limited, and hence, the likelihood of light exiting the surface, and the likelihood of light being re-injected into the surface, is difficult to control or predict.